CN117725149A - Query method, device and equipment for content feedback - Google Patents
Query method, device and equipment for content feedback Download PDFInfo
- Publication number
- CN117725149A CN117725149A CN202211104824.4A CN202211104824A CN117725149A CN 117725149 A CN117725149 A CN 117725149A CN 202211104824 A CN202211104824 A CN 202211104824A CN 117725149 A CN117725149 A CN 117725149A
- Authority
- CN
- China
- Prior art keywords
- content
- query
- search
- complaint
- data
- Prior art date
- Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
- Pending
Links
- 238000000034 method Methods 0.000 title claims abstract description 65
- 239000013598 vector Substances 0.000 claims abstract description 72
- 238000012545 processing Methods 0.000 claims abstract description 46
- 238000011156 evaluation Methods 0.000 claims description 100
- 230000011218 segmentation Effects 0.000 claims description 30
- 230000004044 response Effects 0.000 claims description 21
- 230000008451 emotion Effects 0.000 claims description 16
- 230000008909 emotion recognition Effects 0.000 claims description 9
- 239000002131 composite material Substances 0.000 claims description 3
- 238000010586 diagram Methods 0.000 description 19
- 238000003860 storage Methods 0.000 description 16
- 238000005516 engineering process Methods 0.000 description 14
- 238000000605 extraction Methods 0.000 description 13
- 238000004458 analytical method Methods 0.000 description 12
- 238000004422 calculation algorithm Methods 0.000 description 6
- 238000006243 chemical reaction Methods 0.000 description 6
- 230000008569 process Effects 0.000 description 6
- 238000004891 communication Methods 0.000 description 4
- 238000004590 computer program Methods 0.000 description 4
- 230000004927 fusion Effects 0.000 description 4
- 238000013473 artificial intelligence Methods 0.000 description 3
- 238000013527 convolutional neural network Methods 0.000 description 3
- 230000009193 crawling Effects 0.000 description 3
- 230000006870 function Effects 0.000 description 3
- 238000002372 labelling Methods 0.000 description 3
- 238000013528 artificial neural network Methods 0.000 description 2
- 230000002457 bidirectional effect Effects 0.000 description 2
- 238000004364 calculation method Methods 0.000 description 2
- 230000006835 compression Effects 0.000 description 2
- 238000007906 compression Methods 0.000 description 2
- 230000003287 optical effect Effects 0.000 description 2
- 238000009877 rendering Methods 0.000 description 2
- 238000012216 screening Methods 0.000 description 2
- 239000007787 solid Substances 0.000 description 2
- 238000012549 training Methods 0.000 description 2
- 230000001960 triggered effect Effects 0.000 description 2
- 230000000007 visual effect Effects 0.000 description 2
- 238000013459 approach Methods 0.000 description 1
- 238000013475 authorization Methods 0.000 description 1
- 230000006399 behavior Effects 0.000 description 1
- 230000009286 beneficial effect Effects 0.000 description 1
- 230000005540 biological transmission Effects 0.000 description 1
- 238000005266 casting Methods 0.000 description 1
- 125000004122 cyclic group Chemical group 0.000 description 1
- 230000002354 daily effect Effects 0.000 description 1
- 238000013500 data storage Methods 0.000 description 1
- 238000011161 development Methods 0.000 description 1
- 230000018109 developmental process Effects 0.000 description 1
- 230000003203 everyday effect Effects 0.000 description 1
- 230000010354 integration Effects 0.000 description 1
- 230000003993 interaction Effects 0.000 description 1
- 238000010801 machine learning Methods 0.000 description 1
- 238000007726 management method Methods 0.000 description 1
- 230000007246 mechanism Effects 0.000 description 1
- 230000005055 memory storage Effects 0.000 description 1
- 238000012986 modification Methods 0.000 description 1
- 230000004048 modification Effects 0.000 description 1
- 230000006855 networking Effects 0.000 description 1
- 230000000306 recurrent effect Effects 0.000 description 1
- 238000007670 refining Methods 0.000 description 1
- 230000002441 reversible effect Effects 0.000 description 1
- 238000007619 statistical method Methods 0.000 description 1
- 238000013179 statistical model Methods 0.000 description 1
- 239000013589 supplement Substances 0.000 description 1
Landscapes
- Information Retrieval, Db Structures And Fs Structures Therefor (AREA)
Abstract
The application discloses a query method, a query device and query equipment for content feedback, and relates to the technical field of computers. The method comprises the following steps: judging the type of the acquired query request to obtain the query type, wherein the query request contains search content; if the query type is text search, carrying out accuracy recognition on the query request to obtain the accuracy of the query request; matching the query request based on the accuracy to obtain subjective data and objective data corresponding to the search content; if the query type is picture searching, obtaining a target vector of a target picture, wherein a query request corresponding to the picture searching comprises the target picture; vector retrieval processing is carried out based on the target vector, and subjective data and objective data corresponding to the search content are obtained; and carrying out joint statistics according to the subjective data and the objective data to obtain a target feedback result of the query request. The method can improve the acquisition efficiency and the data accuracy of the feedback condition of the recommended content.
Description
Technical Field
The present disclosure relates to the field of computer technologies, and in particular, to a method, an apparatus, and a device for querying content feedback.
Background
The advent of internet advertising has enabled advertisers to promote goods or services using more advertising approaches. After the advertisement content is put on the advertisement platform, the advertiser or the advertisement platform needs to perform subsequent analysis according to the feedback condition after the advertisement content is put on.
In the related art, an advertiser or an advertisement platform pulls a log table of advertisement content on the advertisement platform, and performs statistical analysis on original data in the log table, so as to determine feedback conditions of the advertisement content, for example, determine exposure rate, click rate, conversion rate and the like of the advertisement content.
However, the process of acquiring the feedback condition of the advertisement content is inefficient, and the feedback data is single.
Disclosure of Invention
The embodiment of the application provides a query method, a query device and query equipment for content feedback, which can improve the acquisition efficiency of feedback conditions of recommended content. The technical scheme is as follows:
in one aspect, a method for querying content feedback is provided, the method comprising:
judging the type of the acquired query request to obtain a query type, wherein the query request comprises search content, and the query type comprises text search and picture search;
If the query type is the text search, performing accuracy recognition on the query request to obtain the accuracy of the query request;
performing matching processing on the query request based on the precision to obtain subjective data and objective data corresponding to the search content, wherein the subjective data are used for indicating the individual evaluation of the candidate recommended content by the account number of the candidate recommended content corresponding to the search content, and the objective data are used for indicating the comprehensive evaluation of the event associated with the candidate recommended content by a designated group;
if the query type is the picture search, acquiring a target vector of a target picture, wherein a query request corresponding to the picture search comprises the target picture;
vector retrieval processing is carried out based on the target vector, and subjective data and objective data corresponding to the search content are obtained;
and carrying out joint statistics according to the subjective data and the objective data to obtain a target feedback result of the query request.
In another aspect, a method for querying content feedback is provided, the method including:
displaying a query interface, wherein the query interface is used for querying recommendation feedback data of recommendation contents in a history period;
Receiving a query operation for searching contents in the query interface, wherein the query operation indicates to match the searching contents, determines candidate recommended contents with matching relation with the searching contents, and acquires objective data and subjective data corresponding to the candidate recommended contents;
displaying target feedback results corresponding to the candidate recommended content; the target feedback result is determined through the objective data and the subjective data corresponding to the candidate recommended content, the objective data is used for indicating comprehensive evaluation of an appointed group on an event associated with the candidate recommended content, and the subjective data is used for indicating individual evaluation of an account number receiving the candidate recommended content on the candidate recommended content.
In another aspect, a query device for content feedback is provided, the device comprising:
the first determining module is used for judging the type of the acquired query request to obtain a query type, wherein the query request comprises search content, and the query type comprises text search and picture search;
the first processing module is used for identifying the accuracy of the query request if the query type is the text search, so as to obtain the accuracy of the query request;
The first processing module is further configured to perform matching processing on the query request based on the accuracy, so as to obtain subjective data and objective data corresponding to the search content, where the subjective data is used to instruct an account number of the candidate recommended content corresponding to the search content to receive individual evaluation of the candidate recommended content, and the objective data is used to instruct a specified group to comprehensively evaluate an event associated with the candidate recommended content;
the second processing module is used for acquiring a target vector of a target picture if the query type is the picture search, wherein the query request corresponding to the picture search comprises the target picture;
the second processing module is further used for carrying out vector retrieval processing based on the target vector to obtain subjective data and objective data corresponding to the search content;
and the second determining module is used for carrying out joint statistics according to the subjective data and the objective data to obtain a target feedback result of the query request.
In another aspect, a query device for content feedback is provided, the device comprising:
the display module is used for displaying a query interface, wherein the query interface is used for querying recommendation feedback data of recommended content in a history period;
The receiving module is used for receiving the query operation of the search content in the query interface, the query operation indicates the matching of the search content, determines candidate recommended content with matching relation with the search content, and acquires objective data and subjective data corresponding to the candidate recommended content;
the display module is also used for displaying target feedback results corresponding to the candidate recommended content; the target feedback result is determined through objective data and subjective data corresponding to the candidate recommended content, the objective data are used for indicating comprehensive evaluation of an appointed group on an event associated with the candidate recommended content, and the subjective data are used for indicating individual evaluation of an account number receiving the candidate recommended content on the candidate recommended content.
In another aspect, a computer device is provided, where the terminal includes a processor and a memory, where the memory stores at least one instruction, at least one program, a code set, or an instruction set, where the at least one instruction, the at least one program, the code set, or the instruction set is loaded and executed by the processor to implement a content feedback query method according to any one of the embodiments of the present application.
In another aspect, a computer readable storage medium is provided, where at least one piece of program code is stored, where the program code is loaded and executed by a processor to implement a content feedback query method according to any one of the embodiments of the present application.
In another aspect, a computer program product or computer program is provided, the computer program product or computer program comprising computer instructions stored in a computer readable storage medium. The processor of the computer device reads the computer instructions from the computer-readable storage medium, and the processor executes the computer instructions, so that the computer device performs the content feedback query method according to any of the above embodiments.
The technical scheme provided by the application at least comprises the following beneficial effects:
when the content feedback condition corresponding to the search content is required to be acquired, determining the query type corresponding to the search content according to the search content indicated by the query request, confirming according to the accuracy corresponding to the query request when the query type corresponding to the search content is determined to be text search, acquiring subjective data and objective data corresponding to the search content based on the accuracy, acquiring a target vector corresponding to the searched picture when the query type corresponding to the search content is determined to be picture search, searching based on the target vector, acquiring subjective data and objective data corresponding to the search content, and carrying out joint statistics through the subjective data and the objective data corresponding to the search content, so that a feedback result corresponding to the query request is obtained. The method realizes the automatic inquiry of the feedback condition of the recommended content through the provided inquiry system, improves the acquisition efficiency of the feedback condition of the recommended content, provides different data feedback according to different inquiry types and accuracy, can improve the accuracy of the feedback result, and simultaneously improves the diversity of the feedback result by combining comprehensive evaluation and individual evaluation, thereby enriching the diversity of data used in downstream result analysis and improving the analysis efficiency of the content feedback condition.
Drawings
In order to more clearly illustrate the technical solutions of the embodiments of the present application, the drawings that are needed in the description of the embodiments will be briefly introduced below, and it is obvious that the drawings in the following description are only some embodiments of the present application, and that other drawings may be obtained according to these drawings without inventive effort for a person skilled in the art.
FIG. 1 is a schematic illustration of an implementation environment provided by an exemplary embodiment of the present application;
FIG. 2 is a flow chart of a method of querying content feedback provided by an exemplary embodiment of the present application;
FIG. 3 is an interface diagram of historical period settings provided by one exemplary embodiment of the present application;
FIG. 4 is a schematic illustration of a display of subjective data provided by an exemplary embodiment of the present application;
FIG. 5 is a flow chart of a method of querying content feedback provided by an exemplary embodiment of the present application;
FIG. 6 is a schematic diagram of a precision query control provided by an exemplary embodiment of the present application;
FIG. 7 is a flow chart of a search service provided by an exemplary embodiment of the present application;
FIG. 8 is a schematic diagram showing the results of target feedback provided by an exemplary embodiment of the present application;
FIG. 9 is a flowchart of a method of querying content feedback provided by an exemplary embodiment of the present application;
FIG. 10 is a schematic illustration of complaint statistics provided by an exemplary embodiment of the present application;
FIG. 11 is a block diagram of a query of advertisement feedback provided by an exemplary embodiment of the present application;
FIG. 12 is a block diagram of a querying device for content feedback provided by an exemplary embodiment of the present application;
FIG. 13 is a block diagram of a querying device for content feedback provided by an exemplary embodiment of the present application;
FIG. 14 is a block diagram of a querying device for content feedback provided by an exemplary embodiment of the present application;
FIG. 15 is a block diagram of a querying device for content feedback provided by an exemplary embodiment of the present application;
fig. 16 is a block diagram of a terminal according to an exemplary embodiment of the present application.
Detailed Description
For the purpose of making the objects, technical solutions and advantages of the present application more apparent, the embodiments of the present application will be described in further detail below with reference to the accompanying drawings.
First, an application scenario of the method of the present application is schematically described, and the method may be applied to at least one of the following scenarios:
first kind: the method is applied to query services of content delivery parties, such as query of advertisement promotion feedback. Schematically, an auditor of an advertiser or an advertisement platform enters a query interface through a first application, wherein the first application is a background data query application provided by the advertisement platform and facing the advertiser or the auditor, and a target feedback result corresponding to the advertisement content is obtained by querying the advertisement content in the query interface. The advertiser can determine whether to continuously bid on the advertisement content according to the advertisement analysis report so as to acquire the put resources, and the auditor can determine whether to put the advertisement content off the shelf in the advertisement platform according to the target feedback result.
Second kind: the method is applied to the query service of the content acquirer. The content acquirer accesses the query interface through the second application, wherein the second application is a content browsing application provided for the content acquirer by the content pushing platform, and queries the appointed content through the query interface, so that the content acquirer obtains the appointed content and simultaneously knows the comprehensive evaluation and the individual evaluation corresponding to the appointed content through the target feedback result corresponding to the appointed content.
It should be noted that the application of the method of the present application is merely illustrative, and the method may also be applied to other query scenarios of content recommendation feedback, which is not limited herein.
Referring to fig. 1, a schematic diagram of an implementation environment provided by an exemplary embodiment of the present application is shown. The computer system of the implementation environment comprises: a terminal device, a server 120 and a communication network 130, wherein the terminal device comprises a first terminal 111 and a second terminal 112.
The terminal equipment comprises various types of equipment such as mobile phones, tablet computers, desktop computers, portable notebook computers, intelligent voice interaction equipment, intelligent household appliances, vehicle-mounted terminals, aircrafts and the like.
The first terminal 111 is running a first application, where the first application provides a feedback query service for content recommendation, and the first user may query, through the first application, feedback conditions corresponding to recommended content put in the content recommendation platform.
The second terminal 112 has a second application running therein, the second application providing a content recommendation service, through which a second user can browse recommended content provided by the content recommendation platform. Illustratively, after the second user receives the recommended content through the second application, the operation of the recommended content by the second user is recorded in the server 120 of the content recommendation platform with sufficient authorization.
The server 120 is configured to provide back-end support for content recommendation services and feedback query services of the content recommendation platform. Illustratively, the recommendation of the content to the second terminal 112 by the server 120 is recorded in a content log. Meanwhile, the server 120 obtains the web page data in the public web page at a designated frequency, and performs sorting analysis according to the web page data to obtain objective data.
When the server 120 receives the query request sent by the first terminal 111, the server 120 matches the search content carried in the query request in the content log, determines objective data and subjective data corresponding to the search content, determines a target feedback result according to the objective data and the subjective data, feeds back the target feedback result to the first terminal 111, and displays the obtained target feedback result in the first terminal 111 on a query interface provided by the first application.
It should be noted that the above-mentioned joint statistics process of objective data and subjective data corresponding to the candidate recommended content may also be implemented by the server 120, which is not limited herein.
Optionally, the first application and the second application may be the same application or different applications provided by the same platform, and when the first application and the second application are the same application, the first application and the second application may be applications logged in on the same application through accounts with different rights, or the second application is a main application, and the first application is a functional module or a functional component on the second application. Alternatively, the first application and the second application may be conventional application software, may be cloud application software, may be implemented as an applet or an application module in a host application program, or may be a web platform, which is not limited herein. Alternatively, the above-described second application may be an electronic commerce application, a short video application, an audio application, a novel application, a map application, or the like, without being particularly limited thereto.
Alternatively, the server 120 may be an independent physical server, a server cluster or a distributed system formed by a plurality of physical servers, or a cloud server that provides cloud services, cloud databases, cloud security, cloud computing, cloud functions, cloud storage, network services, cloud communication, middleware services, domain name services, security services, a content distribution network (Content Delivery Network, CDN), and basic cloud computing services such as big data and an artificial intelligence platform.
Cloud Technology (Cloud Technology) refers to a hosting Technology that unifies serial resources such as hardware, software, network and the like in a wide area network or a local area network to realize calculation, storage, processing and sharing of data. The cloud technology is based on the general names of network technology, information technology, integration technology, management platform technology, application technology and the like applied by the cloud computing business mode, can form a resource pool, and is flexible and convenient as required. Cloud computing technology will become an important support. Background services of technical networking systems require a large amount of computing, storage resources, such as video websites, picture-like websites, and more portals. Along with the high development and application of the internet industry, each article possibly has an own identification mark in the future, the identification mark needs to be transmitted to a background system for logic processing, data with different levels can be processed separately, and various industry data needs strong system rear shield support and can be realized only through cloud computing.
In some embodiments, the server 120 described above may also be implemented as a node in a blockchain system. Blockchain (Blockchain) is a new application mode of computer technologies such as distributed data storage, point-to-point transmission, consensus mechanisms, encryption algorithms, and the like.
The terminal device and the server 120 are illustratively connected through a communication network 130, where the communication network 130 may be a wired network or a wireless network, which is not limited herein.
Referring to fig. 2, a query method of content feedback according to an embodiment of the present application is illustrated, where in the embodiment of the present application, the method is applied to the first terminal 111 shown in fig. 1, and the method includes:
step 210, a query interface is displayed.
Illustratively, the query interface is configured to query recommendation feedback data for recommended content over a historical period of time.
In this embodiment of the present application, the query interface is an application interface provided by the first application. Illustratively, the first application is an application provided by the content recommendation platform to the first user for querying a recommendation feedback scenario of the recommended content. Alternatively, the first user may be a user who performs content delivery on the content recommendation platform, or may be a manager of the content recommendation platform.
Alternatively, the recommended content may be at least one of text content, image content, video content, and audio content.
Alternatively, the history period may be a period set by the first user through the first terminal; alternatively, the above-described history period may be a system-specified period of time.
In some embodiments, when the historical period is a period set by the terminal, a period setting control is provided in the query interface, and the start and stop time of the historical period can be configured through the period setting control. In one example, as shown in fig. 3, which illustrates an interface schematic diagram of a history period setting provided in an exemplary embodiment of the present application, a period setting control 310 is provided in the query interface 300, where the period setting control 310 includes a start time sub-control 311 and an end time sub-control 312, a start time of a queried history period may be configured by the start time sub-control 311, and an end time of a queried history period may be configured by the end time sub-control 312.
Step 220, a query operation for search content in a query interface is received.
Schematically, the above query operation indicates that the search content is matched in the content log in a target query mode, candidate recommended content having a matching relationship with the search content is determined, objective data and subjective data corresponding to the candidate recommended content are obtained, and the target query mode is used for indicating the matching degree between the search content and the candidate recommended content.
Alternatively, the search content may be at least one of text content, picture content, and voice content.
In some embodiments, an input box is provided in the query interface through which text content may be entered. That is, an input operation for text content is received in an input box provided in a query interface; a query operation for text content is received on a query control provided in a query interface.
In some embodiments, a picture upload control is provided in the query interface through which the picture content may be uploaded. Namely, receiving uploading operation of the picture content on a picture uploading control provided in a query interface; a query operation for the picture content is received on a query control provided in a query interface.
In some embodiments, a sound recording control is provided in the query interface through which the voice content may be uploaded. Namely, receiving input operation of voice content on a recording control provided in a query interface; a query operation for voice content is received on a query control provided in a query interface.
Alternatively, when the query operation is triggered by the query control, it may be implemented by at least one operation of a click operation, a long press operation, a double click operation, or the like on the query control.
Alternatively, the above-mentioned query operation may be implemented through a preset shortcut key in addition to the query control, for example, when the first terminal receives an input operation of the "enter" shortcut key, a query operation for searching for content is triggered.
In some embodiments, the range of recommended content recorded in the content log is associated with account rights of a query account logged in the first application. In an example, when the query account logged in the first application is an administrator account, the content recorded in the content log may be all recommended content in the content recommendation platform, or recommended content currently in a put state in the content recommendation platform. In another example, when the query account logged in the first application is a delivery account of the content delivery party, the content recorded in the content log may be recommended content that is historically delivered by the delivery account on the content recommendation platform.
Illustratively, the objective data is used to indicate a comprehensive evaluation of events associated with candidate recommended content by a specified group, which in one example may be hot events in the internet; the subjective data are used for indicating the individual evaluation of the candidate recommended content by the receiving account. The received account is an account for receiving candidate recommended content in a history period.
In some embodiments, the event may be content obtained by the server crawling and post-finishing hot content in the public web page according to a specified frequency; alternatively, the event may be content determined by a search term in a search ranking list obtained by crawling from the public web page by the server. Illustratively, the server sets a corresponding event identifier for the obtained event and correspondingly stores the event identifier in the event database so as to obtain downstream related data.
In some embodiments, to ensure freshness of the event, the server updates the data in the event database according to a specified period.
Illustratively, when determining the event associated with the candidate recommended content, the candidate event may be determined as the event associated with the candidate recommended content by calculating the association degree between the candidate recommended content and the candidate event in the event library, when the association degree between the candidate recommended content and the candidate event reaches the specified association threshold. Alternatively, in determining the degree of association between the candidate recommended content and the candidate event, the determination may be made by calculating the semantic similarity between the candidate recommended content and the candidate event.
In some embodiments, after receiving the query operation, the first application generates a query request according to the query operation, where the query request includes search content, the first application sends the query request to a server, and the server determines candidate recommended content matching the search content from a content log according to the query request.
Illustratively, after determining candidate recommended content corresponding to the search content, the server acquires objective data corresponding to the candidate recommended content from a first database corresponding to the content log, and acquires subjective data corresponding to the candidate recommended content from a second database corresponding to the content log.
In some embodiments, the candidate recommended content corresponds to a content tag, an event corresponding to the candidate recommended content is determined according to the content tag, and corresponding objective data is obtained according to the event. In some embodiments, the objective data is data obtained by analyzing and sorting web content crawled from a public web page by a crawler technology. In some embodiments, when web content acquisition is implemented through crawler technology, in order to acquire web content more specifically, keywords may be set, that is, only web content associated with the keywords in the public web page may be acquired.
In some embodiments, the content identifier corresponds to the candidate recommended content, and the server obtains subjective data corresponding to the candidate recommended content from the second database according to the content identifier, where the subjective data is data obtained in a flow log recorded by the server when the server recommends the content to the second terminal through the second application; and/or the data obtained from the operation data of the receiving account uploaded by the second terminal on the recommended content by the server under the condition that the receiving account is fully authorized.
Optionally, the subjective data includes at least one of recommended content exposure, conversion, click/browse, collection, praise, report, non-interest annotation, and delivery resource consumption. The exposure is used for indicating the times of pushing the recommended content to the receiving account; the conversion amount is used to indicate the number of times the recommended content is clicked and generates a resource conversion (e.g., generates a purchase behavior, etc.); the click/browse amount indicates the number of times the recommended content is clicked/browsed by the receiving account; the collection indicates the number of collection of the account number in which the recommended content is received; the praise amount indicates the number of praise of the recommended content by the receiving account; the reporting amount indicates the number of the recommended content reported by the received account; the non-interest labeling quantity indicates the quantity of the recommended content labeled as non-interest content by the receiving account; the released resource consumption indicates a resource consumed by the recommended content when the content recommendation platform is promoted, for example, a consumption of the released funds.
Illustratively, the target query mode includes an accurate query mode and a fuzzy query mode, wherein the matching degree between the content set corresponding to the candidate recommended content queried by the accurate query mode and the search content is higher than the matching degree between the content set corresponding to the candidate recommended content queried by the fuzzy query mode and the search content. In some embodiments, the plurality of candidate recommended content queried in the fuzzy query mode includes the plurality of candidate recommended content queried in the accurate query mode.
In one example, the above-described settings of the precise query mode and the fuzzy query mode are indicated by a precise query control in the query interface, illustratively, the precise query mode is determined to be enabled when the precise query control is on, and the fuzzy query mode is determined to be enabled when the precise query control is off.
And 230, displaying target feedback results corresponding to the candidate recommended content.
Illustratively, the target feedback result is determined by objective data and subjective data corresponding to the candidate recommended content.
In some embodiments, the target feedback result includes data derived from objective data. Optionally, the target feedback result includes an event corresponding to the candidate recommended content and evaluation content corresponding to the event. Optionally, the evaluation content corresponding to the event may be obtained through crawling statistics, and the evaluation heat in the specified web page reaches the content of the specified heat threshold, for example, a plurality of evaluation contents corresponding to the event in a plurality of public web pages are obtained, the plurality of evaluation contents are ranked according to the number of praise, and the top N evaluation contents with the highest number of praise are displayed in the target feedback result.
In some embodiments, the target feedback result includes data derived from subjective data. Optionally, the target feedback result includes at least one subjective data of exposure, conversion, click/browse, collection, praise, report, non-interest labeling, and resource consumption corresponding to the candidate recommended content.
Alternatively, when subjective data in the target feedback result is displayed, the large-disc feedback data may be displayed in addition to subjective data corresponding to candidate recommended content. The large disc feedback data are used for indicating subjective data corresponding to all recommended contents in the content recommendation platform. For example, the target feedback result may correspond to display of the large disc exposure when displaying the exposure corresponding to the candidate recommended content. Alternatively, subjective feedback data of recommended content under the category corresponding to the candidate recommended content may also be displayed for comparison with subjective data of the candidate recommended content. Optionally, subjective feedback data of recommended content matched with the search content in other content recommendation platforms can be displayed, and it is noted that the subjective feedback data acquired on the other content recommendation platforms are acquired through public data provided by the public platform, and the acquisition and the use of the data are authorized and legal.
In an example, as shown in fig. 4, a schematic display diagram of subjective data provided in an exemplary embodiment of the present application is shown, and a target feedback result obtained by querying the search content 401 is displayed in the query interface 400, where the target feedback result includes first number information 411 of candidate recommended content obtained by matching the search content 401, second number information 412 corresponding to all recommended content in the content recommendation platform, and third number information 413 of matched recommended content corresponding to other platforms; the target feedback result further includes a first exposure 421 corresponding to the candidate recommended content, a second exposure 422 corresponding to the recommended content, and a third exposure 423 corresponding to the matched recommended content corresponding to the other platform; the target feedback result also comprises first released resource consumption information 431 corresponding to the candidate recommended content and second released resource consumption information 432 corresponding to all the recommended content in the content recommendation platform; the target feedback result also includes a first click rate 441 corresponding to the candidate recommended content and a second click rate 442 corresponding to all recommended contents in the content recommendation platform; the target feedback result further includes the first complaint information 451 corresponding to the candidate recommended content and the second complaint information 452 corresponding to all the recommended content in the content recommendation platform.
In some embodiments, the generating process of the target feedback result may be performed by the first terminal, or may be performed by the server, that is, the server performs joint statistics according to objective data and subjective data to obtain the target feedback result, and sends the target feedback result to the first terminal, where the first terminal displays the target feedback result.
In summary, in the content feedback query method provided in the embodiment of the present application, when the feedback condition of the recommended content is queried in the query interface, the objective data and the subjective data of the candidate recommended content matched with the searched content are obtained to generate the corresponding target feedback result, that is, the recommended feedback data of the recommended content is queried rapidly through the query interface, so that the acquisition efficiency of the feedback condition of the recommended content is improved, and meanwhile, the comprehensive evaluation and the individual evaluation are combined, so that the diversity of the feedback information of the content recommendation is improved, the diversity of the data used in the downstream result analysis is enriched, and the analysis efficiency of the content recommendation feedback condition is improved.
Referring to fig. 5, a query method of content feedback according to an embodiment of the present application is illustrated, where in the embodiment of the present application, the method is applied to the server 120 shown in fig. 1, and the method includes:
And 510, judging the type of the acquired query request to obtain the query type.
The query request is schematically a request sent by the first terminal, wherein the query request comprises search content, and the search content is used for querying from a content log to obtain candidate recommended content with a matching relationship.
Optionally, the query types include text searches and picture searches. For the query requests of different query types, the server inputs the search content into different query services for processing, and illustratively, if the query type is text search, the search content is input into the text search service, and if the query type is picture search, the search content is input into the picture search service.
Illustratively, the server determines the query type based on the content type corresponding to the search content. Alternatively, the search content may be at least one of text content, picture content, and voice content.
In some embodiments, when it is determined that the search content is text content, it is determined that the query type corresponding to the query request is text search, and the text content is input into a text search service. When the search content is determined to be voice content, determining that the query type corresponding to the query request is text search, illustratively, performing voice-to-text processing on the voice content by the server to obtain text content corresponding to the voice content, and inputting the text content into a text search service.
When the search content is determined to be the picture content, determining that the query type corresponding to the query request is picture search, and inputting the picture content into a picture search service.
In some embodiments, the search content may also be obtained by combining contents of a plurality of different content types, and when the search content includes contents of a plurality of different content types, the server divides the search content according to the content types to obtain a plurality of sub-search contents, and inputs each sub-content into the query service under the query type corresponding to the content type.
And step 520, if the query type is text search, performing accuracy recognition on the query request to obtain the accuracy of the query request.
In the embodiment of the application, different search precision requirements are corresponding to text search. In some embodiments, the query request further includes a query pattern, where the query pattern includes an exact query pattern and a fuzzy query pattern. Illustratively, in response to the query mode being an accurate query mode, determining that the accuracy corresponding to the query request is a first accuracy; responding to the query mode as a fuzzy query mode, and determining the accuracy corresponding to the query request as a second accuracy; wherein the first accuracy is higher than the second accuracy.
In some embodiments, the server is provided with precision query services for precision query patterns and fuzzy query services for fuzzy query patterns.
Illustratively, when the first terminal indicates to query the search content, the first terminal may indicate the query mode, that is, the first terminal may be in the query request according to the setting condition of the user on the query mode.
In one example, as shown in FIG. 6, a schematic diagram of a precision query control 610 provided in one exemplary embodiment of the present application is shown, including an input box 620 and the precision query control 610 in a search area 600 provided in a query interface. When the precision query control 610 is in an on state, it is determined that precision query services are used, i.e., a precision query mode is indicated in the query request; when the precision query control 610 is in the off state, it is determined that a fuzzy query service is used, i.e., a fuzzy query pattern is indicated in the query request.
In some embodiments, the query request carries a state identifier of the precise query control, where the state identifier is used to indicate the precise query control state in the query interface. In one example, the status flag may be a 0/1 flag, and when the status flag is "0", then the precise query control is indicated to be in an off state, i.e., it is determined that the ambiguous query mode is enabled; when the status flag is "1," then the precision query control is indicated as being in an on state, i.e., it is determined that the precision query mode is enabled.
In other embodiments, the application of the precision query pattern or the fuzzy query pattern is determined by the background. Illustratively, when matching is performed in the content log according to the search content, the server starts an accurate query mode, and when the number of candidate recommended content obtained by matching in the accurate query mode is determined to not reach a query threshold, the server starts a fuzzy query mode to supplement the candidate recommended content. Optionally, the query threshold may be preset by the system, or may be set through a query interface. Illustratively, the query results, whether in the precise query mode or in the fuzzy query mode, are displayed in the query interface when the target feedback results are displayed, depending on the query mode ultimately used by the server.
And 530, carrying out matching processing on the query request based on the accuracy to obtain subjective data and objective data corresponding to the search content.
In some embodiments, determining candidate recommended content in the content log in a first matching mode in an accurate query mode, and acquiring subjective data and objective data corresponding to the candidate recommended content in the content log; and determining candidate recommended contents in the content log in a second matching mode by the fuzzy query mode, and acquiring subjective data and objective data corresponding to the candidate recommended contents in the content log. Illustratively, the objective data is used to indicate a comprehensive evaluation of events associated with candidate recommended content by a specified group, which in one example may be hot events in the internet; the subjective data are used for indicating individual evaluation of candidate recommended content by a receiving account, wherein the receiving account is an account for receiving the candidate recommended content.
Illustratively, the first matching means indicates that the search content is matched with the recommended content in the content log as a complete phrase string. That is, in response to determining that the accuracy corresponding to the query request is the first accuracy, the search content is used as a phrase character string and is matched with the recommended content in the content log, and the first matching degree between the search content and the recommended content is determined; determining recommended content with the first matching degree higher than a first matching threshold value as candidate recommended content; and obtaining subjective data and objective data corresponding to the candidate recommended content.
In one example, when determining to query the text content through the precision query service, a character string direct matching manner is used to determine candidate recommended content from the content log. In one example, the exact query sub-module will perform search matching of candidate recommended content using match_prase matching rules in an index provided by a log search (elastic search) based on text content. Among them, the elastic search is a distributed open source search and analysis engine, which is applicable to all types of data including text, numbers, geospatial, structured and unstructured data, etc., and is excellent in terms of speed and scalability, and is also capable of indexing various types of content. It is noted that other matching rules may be used in performing exact query matching, and are illustrated only schematically herein as match_prase.
Illustratively, the second matching means indicates dividing the search content so as to match the search content at a word segmentation granularity. Namely, in response to determining that the precision corresponding to the query request is the second precision, performing word segmentation encoding on at least one candidate word segment to obtain word segmentation encoding representation corresponding to each candidate word segment; determining a first text feature representation corresponding to the search content based on the word segmentation code representation corresponding to the candidate word segmentation; acquiring second text characteristic representations corresponding to each recommended content in the content log; determining a second degree of matching between the search content and the recommended content based on feature similarity between the first text feature representation and the second text feature representation; determining recommended content with the second matching degree higher than a second matching threshold value as candidate recommended content; and obtaining subjective data and objective data corresponding to the candidate recommended content.
Illustratively, when extracting features of the text content, segmenting the text content to obtain at least one candidate segmented word, and encoding each candidate segmented word to obtain a segmented word encoding representation corresponding to each candidate segmented word. When the text content comprises a plurality of segmented words, the segmented word coding representations corresponding to the segmented words can be connected, so that text characteristic representations corresponding to the text content can be obtained.
Optionally, when the text content is segmented, at least one algorithm for realizing word segmentation processing, such as a forward maximum matching word segmentation algorithm, a reverse maximum matching word segmentation algorithm, a bidirectional maximum matching word segmentation algorithm, a minimum segmentation word segmentation algorithm, an N-element statistical model, and the like, may be used.
In one example, at least one candidate word is derived by "barking" (jieba) word segmentation, the at least one candidate word is input to a bi-directional transcoding model (Bidirectional Encoder Representation from Transformers, BERT), and an embedded (empdding) vector is output as text content through a Hidden (Hidden) layer in the BERT model, i.e., a text feature representation of the text content is derived.
In other embodiments, feature extraction may also be performed on the text content according to the word segmentation condition of the text content. Illustratively, the text content is divided into words to obtain at least one candidate word, and each candidate word is encoded to obtain a word-division encoded representation corresponding to each candidate word. When the text content includes a plurality of sub-words, the sub-word coded representations corresponding to the sub-words can be linked to obtain text feature representations corresponding to the text content.
In other embodiments, coding can be performed according to the word segmentation condition and the word segmentation condition of the text content, schematically, word segmentation coding representation corresponding to the text content is obtained based on the word segmentation condition of the text content, word feature fusion is performed through the word segmentation coding representation and the word segmentation coding representation, word fusion representation corresponding to the text content is obtained, and the word fusion representation is used as text feature representation of the text content.
In one example, when word vector encoding is performed on target text content in units of divided words, word vector encoding representation of a first dimension corresponding to the text content is obtained; word vector coding is carried out on the text content by taking word segmentation as a unit, and a first word vector coding representation of a second dimension corresponding to the text content is obtained; converting the first word vector encoded representation of the second dimension into a second word vector encoded representation of the first dimension; and fusing the word vector coding representation and the second word vector coding representation to obtain a word fusion representation. Before encoding the segmented words, copying each segmented word corresponding to the text content, namely, copying each segmented word for N times, wherein N can be the number of segmented words in the text content.
Alternatively, when merging the character vector encoded representation and the second word vector encoded representation, the character vector encoded representation and the second word vector encoded representation may be added, or the character vector encoded representation and the second word vector encoded representation may be multiplied.
Alternatively, the feature similarity may be determined by calculating at least one of a euclidean distance, a cosine distance, a mahalanobis distance, a hamming distance, etc. between the features, which is not limited herein.
In one example, the text feature representation of the text content is input into a facebook AI similarity search (Facebook AI Similarity Search, faiss) service for vector retrieval, and content identifiers corresponding to candidate recommended content are returned.
The candidate recommended contents obtained through the first matching mode form a first content set, the candidate recommended contents obtained through the second matching mode form a second content set, and the matching degree between the first content set and the search content is higher than that between the second content set and the search content.
In step 540, if the query type is picture search, a target vector of the target picture is obtained.
The query request corresponding to the picture search comprises a target picture. Illustratively, the server is provided with a picture searching service, and the picture searching service performs picture coding on the searched content to obtain a target vector of the target picture.
Schematically, the target image is input into an image feature extraction network obtained through pre-training, and the target vector is obtained through output. Alternatively, the image feature extraction network may be at least one of a visual geometry group network (Visual Geometry Group Network, VGGNet), a depth residual network (Deep residual network, resNet), a convolutional neural network (Convolutional Neural Networks, CNN), a cyclic neural network (Recurrent Neural Network, RNN), a transducer, and the like, which may perform image feature extraction.
In one example, the image feature extraction network may be a VGG16 model, that is, the target image is input into the VGG16 model, and the image casting vector is output.
And 550, carrying out vector retrieval processing based on the target vector to obtain subjective data and objective data corresponding to the searched content.
Schematically, an image vector corresponding to recommended content in a content log is obtained; determining at least one candidate recommended content from recommended content in a content log through feature similarity between the target vector and the image vector; subjective data and objective data corresponding to candidate recommended content are obtained; wherein the at least one candidate recommended content comprises at least one of picture content and video content. That is, the picture search service implements a function of searching for pictures in a picture or searching for videos in a picture.
In some embodiments, the image information corresponding to the recommended content in the content log is subjected to feature extraction processing performed in advance to obtain an image vector corresponding to the recommended content, and the image vector corresponding to the recommended content is correspondingly stored. And when the graph search is carried out, carrying out feature similarity calculation on the target vector of the target picture and the image vector corresponding to the recommended content, and determining the recommended content as candidate recommended content in response to the feature similarity between the target vector and the image vector reaching a similarity threshold. Alternatively, the above-mentioned similarity may be determined by calculating at least one of a euclidean distance, a cosine distance, a mahalanobis distance, a hamming distance, etc. between the features, which is not limited herein.
In one example, the target vector is input to a Faiss service for vector retrieval, and a content identifier corresponding to the candidate recommended content is returned.
In other embodiments, in the process of querying candidate recommended content for the target picture, in addition to querying the recommended content in the form of an image, querying the recommended content in the form of a video may be performed. In some embodiments, image frames corresponding to recommended content in the form of video are subjected to feature extraction processing to obtain picture vectors for characterizing the video. Optionally, when extracting features for image frames of the video, coding all the image frames in the video, and connecting the obtained picture feature coded representations to obtain coded representations representing features of the video; alternatively, to reduce the amount of data corresponding to a picture-encoded representation of a video, key frames in the video may be selected to be encoded to obtain picture vectors that characterize the video.
Illustratively, after determining candidate recommended content corresponding to the search content, the server acquires objective data corresponding to the candidate recommended content from a first database corresponding to the content log, and acquires subjective data corresponding to the candidate recommended content from a second database corresponding to the content log.
Referring to fig. 7, a flowchart of a search service provided in an exemplary embodiment of the present application is shown, where the flowchart includes: 710, front end inputs search content; 720, judging the search type, if text search is performed, executing 731, if picture search is performed, executing 741;731, judging whether to accurately inquire, if yes, executing 732, and if not, executing 733;732, performing search matching using phrase matching (match_phrase) in the log search; 733, "crust" is used in the word; 734, obtaining hidden layer output as sentence embedding (embedding) vectors based on the BERT model; 735 vector retrieval based on Faiss; 741, obtaining a picture embedding (embedding) vector by using a VGG16 model; 742, searching for pictures or searching for videos based on Faiss; and 750, acquiring the related data of the candidate recommended content after determining the candidate recommended content.
And step 560, carrying out joint statistics according to the subjective data and the objective data to obtain a target feedback result of the query request.
The objective data is used for indicating comprehensive evaluation of the appointed group on the event associated with the candidate recommended content, the subjective data is used for indicating individual evaluation of the candidate recommended content by the receiving account, and the receiving account is an account for receiving the candidate recommended content in a historical period.
In some embodiments, the server includes a statistics service, and after the candidate recommended content and the corresponding objective data and subjective data are input into the statistics service, the statistics service performs statistics according to a plurality of specified indexes, so as to obtain a target feedback result.
In some embodiments, the target feedback result includes data derived from objective data. Optionally, the target feedback result includes an event corresponding to the candidate recommended content and evaluation content corresponding to the event.
In some embodiments, the target feedback result includes data derived from subjective data. Optionally, the target feedback result includes at least one subjective data of exposure, conversion, click/browse, collection, praise, report, non-interest labeling, and resource consumption corresponding to the candidate recommended content.
In some embodiments, industry category information corresponding to the query result is also displayed in the target feedback result. Illustratively, the industry type information is used for indicating the industry category to which the candidate recommended content determined by the query according to the search content belongs. Optionally, when the industry category information is displayed, industry categories corresponding to all candidate recommended contents can be selected to be displayed, that is, a plurality of industry categories can be displayed in the industry category information, and the industry categories can be ranked according to the ratio of the number of the corresponding candidate recommended contents in the total number of the candidate recommended contents; alternatively, when the industry category information is displayed, only the industry category with the largest number of candidate recommended contents may be displayed, that is, only the industry category with the largest number of candidate recommended contents may be displayed in the industry category information.
In some embodiments, the candidate recommended content may also be displayed in the target feedback result. Alternatively, a specified number of partial candidate recommended contents may be displayed in consideration of the larger number of the retrieved candidate recommended contents.
In some embodiments, a specified number of candidate recommended content meeting the screening criteria may be displayed in the target feedback result. The above screening conditions may be preset by the system or may be terminal-customized. In one example, candidate recommended content that shows the highest click through may be selected, N being a positive integer.
Referring to fig. 8, a schematic display diagram of a target feedback result provided by an exemplary embodiment of the present application is shown, in which a target feedback result obtained by querying a search content 801 is displayed in a query interface 800, and industry category information 810 corresponding to candidate recommended content obtained by matching the search content 801 and five display recommended content 820 with the highest resource consumption in the candidate recommended content are displayed in the target feedback result.
In summary, in the content feedback query method provided in the embodiment of the present application, when a content feedback condition corresponding to a search content needs to be obtained, a query type corresponding to the search content is determined according to the search content indicated by the query request, when the query type corresponding to the search content is determined to be text search, the query is confirmed according to the accuracy corresponding to the query request, subjective data and objective data corresponding to the search content are obtained based on the accuracy, when the query type corresponding to the search content is determined to be picture search, a target vector corresponding to the searched picture is obtained, search is performed based on the target vector, subjective data and objective data corresponding to the search content are obtained, and joint statistics is performed through the subjective data and the objective data corresponding to the search content, so as to obtain a feedback result corresponding to the query request. The method realizes the automatic inquiry of the feedback condition of the recommended content through the provided inquiry system, improves the acquisition efficiency of the feedback condition of the recommended content, provides different data feedback according to different inquiry types and accuracy, can improve the accuracy of the feedback result, and simultaneously improves the diversity of the feedback result by combining comprehensive evaluation and individual evaluation, thereby enriching the diversity of data used in downstream result analysis and improving the analysis efficiency of the content feedback condition.
Referring to fig. 9, a query method for content feedback provided in an exemplary embodiment of the present application is shown, in this embodiment of the present application, subjective data further includes complaint information of candidate recommended content by an account, where steps 941 to 943 are performed after step 230, and the method includes:
step 941, a complaint statistical result corresponding to the complaint information is displayed in the target feedback result, and the complaint statistical result includes a content list corresponding to the complaint content.
In the embodiment of the application, the complaint statistical result is obtained by counting complaint information of candidate recommended contents. Illustratively, when the user receives the recommended content pushed by the content recommendation platform through the second terminal, the user may complain about the received recommended content.
The complaint statistical result comprises at least one data of total complaint quantity, complaint rate, complaint flow distribution data, complaint industry distribution information, complaint reason distribution data and complaint content. The complaint quantity is used for indicating the times of receiving the complaints of the user in the candidate recommended content; the complaint rate is used to indicate a ratio between the number of complaints of the candidate recommended content and a specified exposure, for example, a complaint rate of millions of exposures; the complaint flow distribution data is used for indicating recommendation paths of candidate recommended contents, for example, a content recommendation platform comprises a social area, an electronic commerce area, an office area and the like, and the complaint flow distribution data is used for indicating the complaint amount received when content is recommended in different areas; the complaint reason distribution data is used for indicating the complaint reason corresponding to the complaint content.
In some embodiments, when the complaint statistics corresponding to the candidate recommended content are displayed, large disc complaint data may also be correspondingly displayed. The large disc feedback data are used for indicating complaint information corresponding to all recommended contents in the content recommendation platform.
In an example, as shown in fig. 10, a schematic diagram of a complaint statistical result provided in an exemplary embodiment of the present application is shown, and a target feedback result obtained by querying search content is displayed in a query interface 1000, where the target feedback result includes a complaint statistical result corresponding to candidate recommended content. The complaint statistics include the total complaint amount 1010, the complaint rate 1020, the complaint browsing distribution data 1030, the complaint industry distribution information 1040, the complaint reason distribution data 1050, and five complaint cases 1060 in the candidate recommended content.
In response to receiving a selection operation of at least one complaint content from the content list, step 942 displays a composite rating corresponding to the at least one complaint content and a degree of association between the at least one complaint content and the event.
In some embodiments, complaint content in the content list may be indicated by a content title and content identification.
In some embodiments, when the number of complaint contents is large, only a part of the complaint contents may be displayed in the content list, and in one example, complaint contents in which the number of complaints in all the complaint contents reaches a specified complaint threshold are displayed in the content list.
Illustratively, feature extraction processing is performed on the complaint content to obtain a first feature representation corresponding to the complaint content. The second feature representation corresponding to the event can be obtained by feature extraction of the evaluation content corresponding to the event. Optionally, the multiple evaluation contents corresponding to the event can be ranked according to the number of praise, and the evaluation content with the highest praise number is subjected to feature extraction to obtain a second feature representation corresponding to the event; or, by acquiring a plurality of evaluation contents corresponding to the event, performing feature extraction on the plurality of evaluation contents respectively to obtain evaluation feature representations corresponding to the evaluation contents, clustering the evaluation feature representations corresponding to the plurality of evaluation contents respectively to obtain cluster clusters corresponding to the plurality of evaluation feature representations, and determining the evaluation feature representation serving as a cluster center in the cluster clusters as a second feature representation corresponding to the event.
After the first feature representation of the complaint content and the second feature representation of the event are obtained, a feature similarity between the first feature representation and the second feature representation is determined as a degree of association between the complaint content and the event. Alternatively, the above-mentioned similarity may be determined by calculating at least one of a euclidean distance, a cosine distance, a mahalanobis distance, a hamming distance, etc. between the features, which is not limited herein.
Step 943, in response to the degree of association between the complaint content and the event reaching the degree of association threshold, the off-shelf control is displayed.
The off-shelf control is used for deleting complaint content from a recommended content log corresponding to the recommendation platform.
Optionally, the association threshold may be preset by the system or may be terminal-customized, which is not limited herein.
The method includes the steps that a trigger operation is received by a shelf-off control, a shelf-off request is sent to a server, the shelf-off request comprises a content identifier corresponding to complained content, after the shelf-off request is received by the server, the server inquires from a content log according to the content identifier to obtain corresponding recommended content, and the recommended content is deleted from the content log, wherein the recommended content stored in the content log is used for content recommendation to a second terminal by a content recommendation platform.
In some embodiments, when the complaint content is off-shelf according to the degree of association between the complaint content and the event, it may be considered whether the positive or negative emotion is conveyed by the composite rating corresponding to the event. Schematically, in response to the degree of association between the complaint content and the event reaching a threshold degree of association, acquiring comprehensive evaluation content corresponding to the event, inputting the comprehensive evaluation content into a text emotion recognition model for emotion recognition, determining an evaluation emotion corresponding to the comprehensive evaluation content, and in response to the evaluation emotion being a negative emotion, displaying the off-shelf control.
The comprehensive evaluation content is obtained by extracting a plurality of evaluation contents corresponding to the event. Alternatively, the evaluation content with the highest heat degree among the plurality of evaluation contents corresponding to the event may be determined as the comprehensive evaluation content of the event; alternatively, a score corresponding to the evaluation content may be calculated according to the account weight of the distribution account corresponding to the evaluation content and the heat corresponding to the evaluation content, and the evaluation content may be ranked according to the score, so as to obtain an evaluation content list, and the evaluation content with the highest score in the evaluation content list may be used as the integrated evaluation content. In one example, when the published account is an official media account, the corresponding account weight is higher than the account weight corresponding to the personal account, and in the official media account of different levels, the account weight of the official media account can be set according to the corresponding official level, which is not limited herein.
Illustratively, the text emotion recognition model is obtained through sample data pre-training, and can be used for semantically classifying positive emotion or negative emotion expressed by the text.
In summary, in the content feedback query method provided in the embodiment of the present application, when the feedback condition of the recommended content is queried in the query interface, the objective data and the subjective data of the candidate recommended content matched with the searched content are obtained to generate the corresponding target feedback result, that is, the recommended feedback data of the recommended content is queried rapidly through the query interface, so that the acquisition efficiency of the feedback condition of the recommended content is improved, and meanwhile, the comprehensive evaluation and the individual evaluation are combined, so that the diversity of the feedback information of the content recommendation is improved, the diversity of the data used in the downstream result analysis is enriched, and the analysis efficiency of the content recommendation feedback condition is improved.
In some embodiments, the server performs offline updating on the data in the first database and the second database corresponding to the content log through the data offline updating module. Optionally, the data offline update module updates the data in the first database and the second database at a preset frequency, in one example, at a daily update frequency.
Referring to fig. 11, a schematic diagram of a module distribution of an advertisement feedback query provided in an exemplary embodiment of the present application is shown, which includes a data offline update module 1110 and a user real-time query module 1120, where the data offline update module 1110 obtains objective data 1111 and subjective data 1112 with a day update frequency, and the subjective data 1112 includes advertisement placement data 1113 and advertisement complaint data 1114. The objective data 1111 and the subjective data 1112 collected every day are input to the Elastic database and the Faiss service 1101 to update the data of the Elastic database and the Faiss service 1101. The user real-time query module 1120 includes a query input module 1121, an Elastic data search and Faiss vector search service 1122, and a data display module 1123, wherein the query input module 1121 inputs the received search content to the Elastic data search and Faiss vector search service 1122, the Elastic data search and Faiss vector search service 1122 outputs a target feedback result to the data display module 1123, and the data display module 1123 displays the target feedback result, wherein the Elastic data search and Faiss vector search service 1122 is updated by the Elastic database and the Faiss service 1101.
Referring to fig. 12, a block diagram of a query device for content feedback according to an exemplary embodiment of the present application is shown, where the device includes the following modules:
the first determining module 1210 is configured to perform type judgment on an obtained query request to obtain a query type, where the query request includes search content, and the query type includes text search and picture search;
the first processing module 1220 is configured to identify, if the query type is the text search, the accuracy of the query request, so as to obtain the accuracy of the query request;
the first processing module 1220 is further configured to perform a matching process on the query request based on the accuracy, so as to obtain subjective data and objective data corresponding to the search content, where the subjective data is used to instruct an account number of the candidate recommended content corresponding to the search content to receive an individual evaluation of the candidate recommended content, and the objective data is used to instruct a specified group to comprehensively evaluate an event associated with the candidate recommended content;
a second processing module 1230, configured to obtain a target vector of a target picture if the query type is the picture search, where the query request corresponding to the picture search includes the target picture;
The second processing module 1230 is further configured to perform vector retrieval processing based on the target vector, so as to obtain subjective data and objective data corresponding to the search content;
and a second determining module 1240, configured to perform joint statistics according to the subjective data and the objective data, so as to obtain a target feedback result of the query request.
In some optional embodiments, the query request further includes a query pattern;
the first processing module 1220 is further configured to determine, in response to the query mode being an accurate query mode, that the precision corresponding to the query request is a first precision; responding to the query mode as a fuzzy query mode, and determining the precision corresponding to the query request as a second precision; wherein the first precision is higher than the second precision.
In some alternative embodiments, as shown in fig. 13, the first processing module 1220 further includes:
a first matching unit 1221, configured to, in response to determining that the precision corresponding to the query request is a first precision, match the search content as a phrase string with the recommended content in the content log, and determine a first degree of matching between the search content and the recommended content;
A first determining unit 1222 for determining the recommended content with the first matching degree higher than a first matching threshold value as a candidate recommended content;
the first obtaining unit 1223 is configured to obtain subjective data and objective data corresponding to the candidate recommended content.
In some optional embodiments, the first determining unit 1222 is further configured to perform word segmentation processing on the search content to obtain at least one candidate word segment corresponding to the search content in response to determining that the precision corresponding to the query request is the second precision;
the first determining unit 1222 is further configured to perform word segmentation encoding on the at least one candidate word segment to obtain a word segmentation encoding representation corresponding to each candidate word segment;
the first determining unit 1222 is further configured to determine a first text feature representation corresponding to the search content based on the word segmentation encoding representation corresponding to the candidate word segmentation;
the first obtaining unit 1223 is further configured to obtain a second text feature representation corresponding to each recommended content in the content log;
the first determining unit 1222 is further configured to determine a second matching degree between the search content and the recommended content based on feature similarity between the first text feature representation and the second text feature representation;
The first determining unit 1222 is further configured to determine the recommended content with the second matching degree higher than a second matching threshold as a candidate recommended content;
the first obtaining unit 1223 is further configured to obtain subjective data and objective data corresponding to the candidate recommended content.
In some alternative embodiments, the second processing module 1230 further comprises:
a second obtaining unit 1231, configured to obtain an image vector corresponding to the recommended content in the content log;
a second determining unit 1232 configured to determine, by feature similarity between the target vector and the image vector, at least one candidate recommended content from recommended content in the content log;
the second obtaining unit 1231 is configured to obtain subjective data and objective data corresponding to the candidate recommended content;
wherein the at least one candidate recommended content comprises at least one of picture content and video content.
Referring to fig. 14, a block diagram of a query device for content feedback according to an exemplary embodiment of the present application is shown, where the device includes the following modules:
the display module 1410 is configured to display a query interface, where the query interface is configured to query recommendation feedback data of recommended content in a history period;
A receiving module 1420, configured to receive a query operation on search content in the query interface, where the query operation indicates matching the search content, determine candidate recommended content having a matching relationship with the search content, and obtain objective data and subjective data corresponding to the candidate recommended content;
the display module 1410 is further configured to display a target feedback result corresponding to the candidate recommended content; the target feedback result is determined through objective data and subjective data corresponding to the candidate recommended content, the objective data are used for indicating comprehensive evaluation of an appointed group on an event associated with the candidate recommended content, and the subjective data are used for indicating individual evaluation of an account number receiving the candidate recommended content on the candidate recommended content.
In some optional embodiments, the subjective data includes complaint information of the candidate recommended content by the receiving account;
the display module 1410 is further configured to display a complaint statistical result corresponding to the complaint information in the target feedback result, where the complaint statistical result is obtained by counting the complaint information of the candidate recommended content;
The complaint statistical result comprises at least one data of total complaint quantity, complaint rate, complaint flow distribution data, complaint industry distribution information, complaint reason distribution data and complaint content.
In some optional embodiments, the complaint statistics result includes a content list corresponding to the complaint content;
the display module 1410 is further configured to display, in response to receiving a selection operation of at least one complaint content from the content list, the comprehensive evaluation corresponding to the at least one complaint content, and a degree of association between the at least one complaint content and the event.
In some optional embodiments, the display module 1410 is further configured to display an off-shelf control, where the off-shelf control is configured to delete the complaint content from a recommended content log corresponding to a recommendation platform, in response to a degree of association between the complaint content and the event reaching a degree of association threshold.
In some alternative embodiments, as shown in fig. 15, the apparatus further comprises:
the obtaining module 1430 is further configured to obtain comprehensive evaluation content corresponding to the event in response to the association degree between the complaint content and the event reaching an association degree threshold, where the comprehensive evaluation content is an evaluation content obtained by refining a plurality of evaluation contents corresponding to the event;
The determining module 1440 is configured to input the comprehensive evaluation content into a text emotion recognition model for performing emotion recognition, and determine an evaluation emotion corresponding to the comprehensive evaluation content, where the text emotion recognition model is configured to perform semantic classification on positive emotion or negative emotion expressed by a text;
the display module 1410 is further configured to display the off-shelf control in response to the assessed emotion being a negative emotion.
In some alternative embodiments, the determining of the association between the complaint content and the event is performed by a determining module 1440, where the determining module 1440 further includes:
an extracting unit 1441, configured to perform feature extraction on the complaint content to obtain a first feature representation corresponding to the complaint content;
an acquisition unit 1442, configured to acquire a plurality of evaluation contents corresponding to the event;
the extracting unit 1441 is further configured to perform feature extraction on the plurality of evaluation contents, so as to obtain an evaluation feature representation corresponding to the evaluation content;
a clustering unit 1443, configured to cluster the evaluation feature representations corresponding to the plurality of evaluation contents respectively, so as to obtain a cluster corresponding to the plurality of evaluation feature representations;
A determining unit 1444, configured to determine an evaluation feature representation serving as a cluster center in the cluster as a second feature representation corresponding to the event;
the determining unit 1444 is further configured to determine a feature similarity between the first feature representation and the second feature representation as a degree of association between the complaint content and the event.
In some alternative embodiments, the search content includes at least one of text content, picture content, and voice content;
the receiving module 1420 is further configured to receive an input operation on the text content in an input box provided in the query interface; receiving a query operation on the text content on a query control provided in the query interface; or receiving uploading operation of the picture content on a picture uploading control provided in the query interface; receiving query operation on the picture content on a query control provided in the query interface; or receiving input operation of the voice content on a recording control provided in the query interface; and receiving a query operation on the voice content on a query control provided in the query interface.
In some optional embodiments, the query interface includes an accurate query control, where the accurate query control is used to set a query mode corresponding to the search content;
the determining module 1440 is further configured to determine a control state of the precise query control in response to receiving the query operation;
the apparatus further comprises: a sending module 1450, configured to send a first query request to a server in response to a control state of the precise query control being in an on state, where the first query request includes the search content and the precise query mode; responding to the control state of the accurate query control in a closed state, and sending a second query request to a server, wherein the second query request comprises the search content and a fuzzy query mode;
the server determines the content set of the candidate recommended content by using a first matching mode according to the accurate query mode, and determines the content set of the candidate recommended content by using a second matching mode according to the fuzzy query mode, wherein the matching degree between the content set corresponding to the first matching mode and the search content is higher than that between the second content set corresponding to the second matching mode and the search content.
It should be noted that: the query device for content feedback provided in the above embodiment is only exemplified by the division of the above functional modules, and in practical application, the above functional allocation may be performed by different functional modules according to needs, that is, the internal structure of the device is divided into different functional modules, so as to complete all or part of the functions described above. In addition, the content feedback query device provided in the above embodiment and the content feedback query method embodiment belong to the same concept, and detailed implementation processes of the content feedback query device are shown in the method embodiment, which is not repeated here.
Fig. 16 shows a block diagram of a terminal 1600 according to an exemplary embodiment of the present application. The terminal 1600 may be: a smart phone, a tablet computer, a dynamic video expert compression standard audio layer 3 player (Moving Picture Experts Group Audio Layer III, MP 3), a dynamic video expert compression standard audio layer 4 (Moving Picture Experts Group Audio Layer IV, MP 4) player, a notebook computer, or a desktop computer. Terminal 1600 may also be referred to by other names of user devices, portable terminals, laptop terminals, desktop terminals, etc.
In general, terminal 1600 includes: a processor 1601, and a memory 1602.
Processor 1601 may include one or more processing cores, such as a 4-core processor, an 8-core processor, and the like. The processor 1601 may be implemented in hardware in at least one of digital signal processing (Digital Signal Processing, DSP), field programmable gate array (Field-Programmable Gate Array, FPGA), programmable logic array (Programmable Logic Array, PLA). The processor 1601 may also include a host processor, which is a processor for processing data in an awake state, also referred to as a central processor (Central Processing Unit, CPU), and a coprocessor; a coprocessor is a low-power processor for processing data in a standby state. In some embodiments, the processor 1601 may be integrated with an image processor (Graphics Processing Unit, GPU) for use in connection with rendering and rendering of content to be displayed by the display screen. In some embodiments, the processor 1601 may also include an artificial intelligence (Artificial Intelligence, AI) processor for processing computing operations related to machine learning.
Memory 1602 may include one or more computer-readable storage media, which may be non-transitory. Memory 1602 may also include high-speed random access memory, as well as non-volatile memory, such as one or more magnetic disk storage devices, flash memory storage devices. In some embodiments, a non-transitory computer readable storage medium in memory 1602 is used to store at least one instruction for execution by processor 1601 to implement a query method of content feedback provided by a method embodiment in the present application.
Illustratively, terminal 1600 also includes other components, and those skilled in the art will appreciate that the structure shown in FIG. 16 is not limiting of terminal 1600 and may include more or fewer components than shown, or may combine certain components, or employ a different arrangement of components.
Those of ordinary skill in the art will appreciate that all or part of the steps in the various methods of the above embodiments may be implemented by a program for instructing related hardware, and the program may be stored in a computer readable storage medium, which may be a computer readable storage medium included in the memory of the above embodiments; or may be a computer-readable storage medium, alone, that is not incorporated into the terminal. The computer readable storage medium stores at least one instruction, at least one program, a code set, or an instruction set, where the at least one instruction, the at least one program, the code set, or the instruction set is loaded and executed by the processor to implement the content feedback query method according to any of the above embodiments.
Alternatively, the computer-readable storage medium may include: read Only Memory (ROM), random access Memory (RAM, random Access Memory), solid state disk (SSD, solid State Drives), or optical disk, etc. The random access memory may include resistive random access memory (ReRAM, resistance Random Access Memory) and dynamic random access memory (DRAM, dynamic Random Access Memory), among others. The foregoing embodiment numbers of the present application are merely for describing, and do not represent advantages or disadvantages of the embodiments.
It will be understood by those skilled in the art that all or part of the steps for implementing the above embodiments may be implemented by hardware, or may be implemented by a program for instructing relevant hardware, where the program may be stored in a computer readable storage medium, and the storage medium may be a read-only memory, a magnetic disk or an optical disk, etc.
The foregoing description of the preferred embodiments is merely exemplary in nature and is in no way intended to limit the invention, since it is intended that all modifications, equivalents, improvements, etc. that fall within the spirit and scope of the invention.
Claims (15)
1. A method for querying content feedback, the method comprising:
judging the type of the acquired query request to obtain a query type, wherein the query request comprises search content, and the query type comprises text search and picture search;
if the query type is the text search, performing accuracy recognition on the query request to obtain the accuracy of the query request;
performing matching processing on the query request based on the precision to obtain subjective data and objective data corresponding to the search content, wherein the subjective data are used for indicating the individual evaluation of the candidate recommended content by the account number of the candidate recommended content corresponding to the search content, and the objective data are used for indicating the comprehensive evaluation of the event associated with the candidate recommended content by a designated group;
If the query type is the picture search, acquiring a target vector of a target picture, wherein a query request corresponding to the picture search comprises the target picture;
vector retrieval processing is carried out based on the target vector, and subjective data and objective data corresponding to the search content are obtained;
and carrying out joint statistics according to the subjective data and the objective data to obtain a target feedback result of the query request.
2. The method of claim 1, wherein the query request further comprises a query pattern;
the performing precision recognition on the query request to obtain the precision of the query request includes:
responding to the query mode as an accurate query mode, and determining the accuracy corresponding to the query request as a first accuracy;
responding to the query mode as a fuzzy query mode, and determining the precision corresponding to the query request as a second precision;
wherein the first precision is higher than the second precision.
3. The method according to claim 2, wherein the matching the query request based on the accuracy to obtain subjective data and objective data corresponding to the search content includes:
Responding to the fact that the precision corresponding to the query request is determined to be the first precision, using the search content as a phrase character string, matching with recommended content in a content log, and determining the first matching degree between the search content and the recommended content;
determining the recommended content with the first matching degree higher than a first matching threshold value as candidate recommended content;
and obtaining subjective data and objective data corresponding to the candidate recommended content.
4. The method according to claim 2, wherein the matching the query request based on the accuracy to obtain subjective data and objective data corresponding to the search content includes:
responding to the determination that the precision corresponding to the query request is the second precision, performing word segmentation processing on the search content to obtain at least one candidate word segment corresponding to the search content;
performing word segmentation coding on the at least one candidate word segment to obtain word segmentation coding representation corresponding to each candidate word segment;
determining a first text feature representation corresponding to the search content based on the word segmentation code representation corresponding to the candidate word segmentation;
acquiring second text characteristic representations corresponding to each recommended content in the content log;
Determining a second degree of matching between the search content and the recommended content based on feature similarities between the first text feature representation and the second text feature representation;
determining the recommended content with the second matching degree higher than a second matching threshold value as candidate recommended content;
and obtaining subjective data and objective data corresponding to the candidate recommended content.
5. A method according to any one of claims 1 to 3, wherein the performing vector retrieval processing based on the target vector to obtain subjective data and objective data corresponding to the search content includes:
acquiring an image vector corresponding to recommended content in a content log;
determining at least one candidate recommended content from recommended content in the content log through the feature similarity between the target vector and the image vector;
obtaining subjective data and objective data corresponding to the candidate recommended content;
wherein the at least one candidate recommended content comprises at least one of picture content and video content.
6. A method for querying content feedback, the method comprising:
displaying a query interface, wherein the query interface is used for querying recommendation feedback data of recommendation contents in a history period;
Receiving a query operation for searching contents in the query interface, wherein the query operation indicates to match the searching contents, determines candidate recommended contents with matching relation with the searching contents, and acquires objective data and subjective data corresponding to the candidate recommended contents;
displaying target feedback results corresponding to the candidate recommended content; the target feedback result is determined through the objective data and the subjective data corresponding to the candidate recommended content, the objective data is used for indicating comprehensive evaluation of an appointed group on an event associated with the candidate recommended content, and the subjective data is used for indicating individual evaluation of an account number receiving the candidate recommended content on the candidate recommended content.
7. The method according to claim 6, wherein the subjective data includes complaint information of the candidate recommended content by the account number of the candidate recommended content;
the displaying the target feedback result corresponding to the candidate recommended content comprises the following steps:
displaying a complaint statistical result corresponding to the complaint information in the target feedback result, wherein the complaint statistical result is obtained by counting the complaint information of the candidate recommended content;
The complaint statistical result comprises at least one data of total complaint quantity, complaint rate, complaint flow distribution data, complaint industry distribution information, complaint reason distribution data and complaint content.
8. The method according to claim 7, wherein the complaint statistics include a content list corresponding to the complaint content;
the method further comprises the steps of:
in response to receiving a selection operation of at least one complaint content from the content list, displaying the comprehensive evaluation corresponding to the at least one complaint content and a degree of association between the at least one complaint content and the event.
9. The method of claim 8, wherein after displaying the composite rating corresponding to the at least one complaint content and the degree of association between the at least one complaint content and the event, further comprising:
and displaying a drop-off control in response to the association degree between the complaint content and the event reaching an association degree threshold, wherein the drop-off control is used for deleting the complaint content from a recommended content log corresponding to a recommendation platform.
10. The method of claim 9, wherein the displaying an off-shelf control in response to the degree of association between the complaint content and the event reaching an association threshold comprises:
Responding to the relation degree between the complaint content and the event reaching a relation degree threshold value, and acquiring comprehensive evaluation content corresponding to the event, wherein the comprehensive evaluation content is obtained by extracting a plurality of evaluation contents corresponding to the event;
inputting the comprehensive evaluation content into a text emotion recognition model for emotion recognition, and determining an evaluation emotion corresponding to the comprehensive evaluation content, wherein the text emotion recognition model is used for carrying out semantic classification on positive emotion or negative emotion expressed by a text;
the off-shelf control is displayed in response to the evaluated emotion being a negative emotion.
11. The method according to any one of claims 8 to 10, wherein the determining of the degree of association between the complaint content and the event comprises:
extracting features of the complaint content to obtain a first feature representation corresponding to the complaint content;
acquiring a plurality of evaluation contents corresponding to the event;
respectively extracting characteristics of the plurality of evaluation contents to obtain evaluation characteristic representations corresponding to the evaluation contents;
clustering the evaluation characteristic representations corresponding to the evaluation contents respectively to obtain clustering clusters corresponding to the evaluation characteristic representations;
Determining the evaluation characteristic representation serving as a clustering center in the cluster as a second characteristic representation corresponding to the event;
feature similarity between the first feature representation and the second feature representation is determined as a degree of association between the complaint content and the event.
12. The method according to any one of claims 6 to 10, wherein the search content includes at least one of text content, picture content, and voice content;
the receiving a query operation for searching content in the query interface comprises the following steps:
receiving an input operation of the text content in an input box provided in the query interface; receiving a query operation on the text content on a query control provided in the query interface;
or,
receiving uploading operation of the picture content on a picture uploading control provided in the query interface; receiving query operation on the picture content on a query control provided in the query interface;
or,
receiving input operation of the voice content on a recording control provided in the query interface; and receiving a query operation on the voice content on a query control provided in the query interface.
13. A content feedback query device, the device comprising:
the first determining module is used for judging the type of the acquired query request to obtain a query type, wherein the query request comprises search content, and the query type comprises text search and picture search;
the first processing module is used for identifying the accuracy of the query request if the query type is the text search, so as to obtain the accuracy of the query request;
the first processing module is further configured to perform matching processing on the query request based on the accuracy, so as to obtain subjective data and objective data corresponding to the search content, where the subjective data is used to instruct an account number of the candidate recommended content corresponding to the search content to receive individual evaluation of the candidate recommended content, and the objective data is used to instruct a specified group to comprehensively evaluate an event associated with the candidate recommended content;
the second processing module is used for acquiring a target vector of a target picture if the query type is the picture search, wherein the query request corresponding to the picture search comprises the target picture;
The second processing module is further used for carrying out vector retrieval processing based on the target vector to obtain subjective data and objective data corresponding to the search content;
and the second determining module is used for carrying out joint statistics according to the subjective data and the objective data to obtain a target feedback result of the query request.
14. A content feedback query device, the device comprising:
the display module is used for displaying a query interface, wherein the query interface is used for querying recommendation feedback data of recommended content in a history period;
the receiving module is used for receiving the query operation of the search content in the query interface, the query operation indicates the matching of the search content, determines candidate recommended content with matching relation with the search content, and acquires objective data and subjective data corresponding to the candidate recommended content;
the display module is also used for displaying target feedback results corresponding to the candidate recommended content; the target feedback result is determined through objective data and subjective data corresponding to the candidate recommended content, the objective data are used for indicating comprehensive evaluation of an appointed group on an event associated with the candidate recommended content, and the subjective data are used for indicating individual evaluation of an account number receiving the candidate recommended content on the candidate recommended content.
15. A computer device comprising a processor and a memory, wherein the memory has stored therein at least one program that is loaded and executed by the processor to implement a method of querying content feedback as claimed in any one of claims 1 to 12.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202211104824.4A CN117725149A (en) | 2022-09-09 | 2022-09-09 | Query method, device and equipment for content feedback |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202211104824.4A CN117725149A (en) | 2022-09-09 | 2022-09-09 | Query method, device and equipment for content feedback |
Publications (1)
| Publication Number | Publication Date |
|---|---|
| CN117725149A true CN117725149A (en) | 2024-03-19 |
Family
ID=90207473
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN202211104824.4A Pending CN117725149A (en) | 2022-09-09 | 2022-09-09 | Query method, device and equipment for content feedback |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN117725149A (en) |
-
2022
- 2022-09-09 CN CN202211104824.4A patent/CN117725149A/en active Pending
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| JP7201730B2 (en) | Intention recommendation method, device, equipment and storage medium | |
| US9449271B2 (en) | Classifying resources using a deep network | |
| US9595053B1 (en) | Product recommendation using sentiment and semantic analysis | |
| CN108733766B (en) | Data query method and device and readable medium | |
| EP2438539B1 (en) | Co-selected image classification | |
| US7860878B2 (en) | Prioritizing media assets for publication | |
| Shi et al. | Learning-to-rank for real-time high-precision hashtag recommendation for streaming news | |
| CN112052387B (en) | Content recommendation method, device and computer readable storage medium | |
| US9069880B2 (en) | Prediction and isolation of patterns across datasets | |
| CN113836131B (en) | Big data cleaning method and device, computer equipment and storage medium | |
| US11126682B1 (en) | Hyperlink based multimedia processing | |
| CN111444357A (en) | Content information determination method and device, computer equipment and storage medium | |
| CN109947902B (en) | Data query method and device and readable medium | |
| CN114238573B (en) | Text countercheck sample-based information pushing method and device | |
| Lubis et al. | A framework of utilizing big data of social media to find out the habits of users using keyword | |
| CN115878841B (en) | Short video recommendation method and system based on improved bald eagle search algorithm | |
| CN111475725A (en) | Method, apparatus, device, and computer-readable storage medium for searching for content | |
| CN112749330A (en) | Information pushing method and device, computer equipment and storage medium | |
| US20240193402A1 (en) | Method and apparatus for determining representation information, device, and storage medium | |
| CN111552884A (en) | Method and apparatus for content recommendation | |
| CN110245357B (en) | Main entity identification method and device | |
| CN106462588B (en) | Content creation from extracted content | |
| CN115168568B (en) | Data content identification method, device and storage medium | |
| CN117132323A (en) | Recommended content analysis method, recommended content analysis device, recommended content analysis equipment, recommended content analysis medium and recommended content analysis program product | |
| CN114662002A (en) | Object recommendation method, medium, device and computing equipment |
Legal Events
| Date | Code | Title | Description |
|---|---|---|---|
| PB01 | Publication | ||
| PB01 | Publication | ||
| SE01 | Entry into force of request for substantive examination | ||
| SE01 | Entry into force of request for substantive examination |